Minimize Mission Performance

SLSTR-A

The SLSTR-A instrument has performed exceptionally well for another year, with all parameters within safe limits. There have been no major anomalies, and only short gaps in data coverage due to ground station issues, manoeuvres or calibration observations.

The cooler has been performing well, with the IR detectors maintained at a stable temperature. Since the cooler cold tip temperature was increased by 1K in July 2018, the time between decontaminations has increased and further measures are planned to increase this further in future.

Radiometric noise levels for the TIR and VIS/SWIR channels have remained stable throughout at pre-launch values. NEDT for the S8 and S9 channels are below 20 mK with no indication of degradation.

Blackbody temperatures have shown a seasonal cycle on top of the daily/orbital temperature cycles, with the highest temperatures of 304.7 K reached during December. The maximum temperature has increased slightly from year to year and will be monitored carefully to prevent it reaching the limit of 305 K necessary to avoid the S7 saturation level.

The VISCAL system is illuminated by the Sun once per orbit and Vicarious calibration results suggest that the system is not degrading significantly over time. The stability is much better than that observed for AATSR on ENVISAT.

The scanners continue to perform well, with orbital mean deviation from the expected position for both nadir and oblique scanners less than 1.5", and a standard deviation less than 5". The flip mirror orbital mean deviation is also less than 1" with a standard deviation <7.4" in the nadir position and <13.6" in oblique position. The worst instantaneous jitter encountered is as good, or better, than previous years.

Level 1 products performance

Validation of the absolute radiometric calibration of the IR channels has been carried out at EUMETSAT using comparisons against IASI-A and B in 2018. The stability of the flight gains, radiometric noise and instrument temperatures suggest that the calibration has not drifted significantly since then.

The VIS and SWIR channels are calibrated via an on-board Solar diffuser-based calibration system. Evaluation of the radiometric calibration has used the techniques developed for AATSR and MERIS and show that the calibration system is stable. Assessment of the VIS channels S1-S3 show good agreement with OLCI and AATSR. At the SWIR wavelengths, there is a significant discrepancy between SLSTR and AATSR and MODIS that must be taken into account in any L2 processing. An adjustment to the L1 processing to correct the main calibration difference is foreseen. The root cause of the anomaly has not been found and is still under investigation.

Geometric calibration is monitored using the GEOCAL tool. Average absolute geometric offsets <0.1 km are achieved for the nadir view and oblique view across-track and <0.2 km for the oblique view along-track.

Several improvements were made in the SLSTR L1 processor from 15th January 2020, including revised ortho-regridding of all channels, revised geo-referencing of SLSTR F1 fire channel, improved geometric calibration for the oblique view, improved S7 brightness temperature upper limit, temporal interpolation of ECMWF meteorological fields, improved quality checks during instrument operations, removal of the SWIR channel ‘C' stripe (time domain integrated map) from the L1 product, improved flags, and update of several NetCDF variable attributes.

Level 2 products performance

The SLSTR-A SL_2_LST product from SLSTR went operational in the Sentinel 3 PDGS on 5th July 2017 with PB 2.16. No additional updates to the retrieval algorithm have been implemented in the IPF since. However, Processing Baseline 2.29 released on 4th April 2018 included the new Probabilistic Cloud Mask implemented in the IPF at Level-1 and carried through to Level-2. Furthermore, from 26th February 2019 an updated ADF of retrieval coefficients has been implemented in PB 2.47, IPF 06.14. Matchups against ten "Gold Standard" in situ stations show that the overall absolute daytime accuracy is 0.81 K and the absolute night-time accuracy is 0.67 K, both of which are within the mission requirements for LST. Comparisons with respect to the operational LSA SAF LST product are within the uncertainty range when considering the uncertainties from the reference products, and thus the products can be interpreted as consistent with each other. Overall, the SL_2_LST product is performing in line with the 1 K mission requirement for LST.

SLSTR-B

Instrument performance

Instrument and blackbody temperatures for SLSTR-B have been stable on top of the daily/orbital and seasonal trends, and consistent with those for SLSTR-A. The cooler has been performing well, with the IR detectors maintained at a stable temperature.

The visible channel radiometric gain shows a variation from orbit to orbit especially in channels S1 and S2. The reason for this behaviour is thought to be due to partial motional chopping of the VIS detectors by an internal aperture in the VIS FPA.  If this is correct, the effect will be present on the earth scene data for S1 and S2.

The NEDT levels are roughly consistent between SLSTR-A and SLSTR-B, except for F1, which shows more orbit-to-orbit variation and higher noise values. This may be caused by motional chopping of the SLSTR-B F1 detectors, which are known to be close to edge of the aperture for SLSTR-B.

The SLSTR-B scanner and flip mean and standard deviations from their expected positions are broadly consistent with SLSTR-A, although the oblique scanner has a slightly larger mean deviation of <3". However, the worst instantaneous jitter for SLSTR-B has increased since the previous year and is higher than SLSTR-A, particularly for the flip mirror.

Level 1 products performance Initial validation of the absolute radiometric calibration of the IR channels has been carried out by EUMETSAT using comparisons against IASI-A and B.  Analysis from the tandem phase comparisons show that the in-flight calibration of SLSTR-B is consistent with that of SLSTR-A.

The S3A and S3B satellites are configured to be 140 degrees out of phase in order to observe complimentary portions of the earth. Figure 1 shows an example combined Level-1 SLSTR-A/SLSTR-B image (daytime only) to show the combined daily SLSTR coverage.

Level 2 products performance

The S3B SL_2_LST product from SLSTR went operational in the Sentinel 3 PDGS on 26th February 2019 with PB 1.19 IPF 06.14. Matchups against ten "Gold Standard" in situ stations show that the overall absolute daytime accuracy is 0.79 K and the absolute night-time accuracy is 0.64 K, both of which are within the mission requirements for LST. As for SLSTR-A, comparisons with respect to the operational LSA SAF LST product are within the uncertainty range when considering the uncertainties from the reference products, and thus the products can be interpreted as consistent with each other. Overall, the SL_2_LST product is performing in line with the 1 K mission requirement for LST.